Feature Extraction
clip
vision
kimihailv commited on
Commit
06c2b25
1 Parent(s): 13811a5

Upload 2 files

Browse files
Files changed (2) hide show
  1. tokenizer.json +0 -0
  2. torch_config.json +26 -0
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
torch_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "text_encoder": {
3
+ "model_type": "bert",
4
+ "dim": 384,
5
+ "context_dim": 768,
6
+ "vocab_size": 250037,
7
+ "padding_idx": 1,
8
+ "num_layers": 12,
9
+ "num_heads": 12,
10
+ "embedding_dim": 256,
11
+ "multimodal_layers_ids": [8, 9, 10, 11],
12
+ "head_one_neuron": true,
13
+ "pooling": "mean",
14
+ "max_position_embeddings": 50,
15
+ "dropout_prob": 0.1
16
+ },
17
+ "image_encoder": {
18
+ "dim": 768,
19
+ "patch_size": 16,
20
+ "image_size": 224,
21
+ "num_layers": 12,
22
+ "num_heads": 12,
23
+ "embedding_dim": 256,
24
+ "pooling": "cls"
25
+ }
26
+ }